Background Endocrine malignancies (including thyroid, adrenal, parathyroid, and pancreatic neuroendocrine tumors) are among the fastest growing cancer diagnoses in the United States, but it is difficult to distinguish benign from malignant tumors by routine clinical, laboratory, and imaging studies. So, even patients who have seemingly benign endocrine tumors often choose to undergo surgery to get a definitive diagnosis in the hopes of ruling out cancer. Most patients with endocrine cancers have a relatively good prognosis. However, anywhere from 10% to 40% (depending on tumor type) have aggressive disease which often cannot be reliably determined at the time of initial treatment. Prognostic markers which can reliably risk stratify patients with high risk of recurrence and death would help determine which patients should receive aggressive initial treatment and close follow up. Furthermore, prognostic markers may also help identify which patients are likely to respond to standard therapy and which patients do not respond to standard therapy if a distinct molecular phenotype is identified. Summary We are making progress with our pan-genomic (mRNA and microRNA expression, copy number changes, and DNA-methylation) analysis of endocrine neoplasms to identify candidate diagnostic and prognostic markers for endocrine malignancies (thyroid, adrenal, neuroendocrine pancreas), and to understand the dysregulated genes/pathways in endocrine cancers. From these analyses, we have identified a miR-30a-LOX axis important in thyroid cancer progression. We found that reduced miR-30a expression and LOX overexpression are significantly associated with more aggressive differentiated thyroid cancer and lower survival rates (even in low-risk conventional papillary thyroid cancer). Our follow up studies have shown that miR-30a-LOX mediates its effect on epithelial-to-mesenchymal transition (EMT) through transcriptionally regulating SNAI2 expression, an EMT transcription factor, and by regulating TIMP4 secretion. Based on these findings, we are currently evaluating the miR-30a-LOX-SNAI2 axis as a prognostic marker in differentiated thyroid cancer. From our integrated analysis of genome-wide gene expression and CpG methylation status in human adrenocortical tumors, we identified candidate tumor suppressor genes that were hypermethylated and downregulated in adrenocortical carcinoma, a common mechanism of tumor suppressor gene silencing in cancer initiation and progression. One of the genes identified is RARRES2, a gene previously shown to mediate an immune response to cancer. Because adrenocortical carcinomas are relatively non-immunogenic, we hypothesized that RARRES2 may have a direct tumor suppressive function. We found that RARRES2 in adrenocortical carcinoma cell lines inhibited growth and invasion in vitro and in vivo in immunocompromised mouse models. RARRES2 was also significantly differentially expressed between adrenal cortical adenoma and carcinoma, and distinguished these tumors with high accuracy. Loss of RARRES2 was seen in both primary and metastatic adrenocortical carcinoma suggesting it is an early event. To understand the mechanism by which RARRES2 functions as a tumor suppressor gene, we investigated its effect on common dysregulated pathways in adrenocortical carcinoma (WNT/beta-catenin, TP53, and IGF). We found that RARRES2 reduced total beta-catenin levels by increasing the phosphorylation and degradation of beta-catenin. In addition to the above promising markers of adrenocortical carcinoma, we have also analyzed and identified urinary metabolites that distinguish between adrenocortical carcinoma and benign adrenocortical tumors.